Linear arrangement
Are animals capable of developing a way of consistently and productively ordering paradigms of form–meaning pairings—in other words, do they have something corresponding to word order in human language?
The data available suggest that the answer is essentially in the affirmative, even if consistent linear arrangement in animals does not seem possible without training. Thus, Premack (1976: 317) notes that chimpanzees can be taught word order, but only with explicit training, and they therefore differ from children, who acquire it on the basis of observational learning. Similar findings have been made on bonobos: The work of Savage-Rumbaugh and Lewin (1994: 160) suggests that acquiring linear arrangements is a matter of systematic training. The bonobo Kanzi ‘‘showed no particular ordering of such symbols during the first month; sometimes he put the action first, sometimes the object. Hide peanut occurred just as often as peanut hide, for instance. But thereafter he began to follow a rather strict order, that of putting the action first and the object second: hide peanut, bite tomato, and so on.’’ The authors suspect that Kanzi learned this arrangement from his caretakers.
But there are also linear arrangements that this bonobo appears to have developed on his own: As we observed above, Savage-Rumbaugh and Lewin (1994: 161) found that Kanzi’s arrangement [action (= lexigram) agent (= gesture)] did not rely on any model that he may have been exposed to. Thus, Kanzi’s arrangement has the appearance of a verb-initial syntax, in that ordering was determined by placement of an action lexigram before an argument, irrespective of the syntactic function of the argument (Greenfield and Savage-Rumbaugh 1990).
Kako (1999: 9) concludes that Kanzi observed a structure that comes close to a word order arrangement: Of the seven major relations involving different types of referents, four showed a statistically significant preference for the following orders: action–agent, entity–demonstrative, goal-action, and object–agent. Furthermore, Kanzi placed his smaller category of gestures (among them three action gestures) almost always after the larger category of lexigrams that he had learned to distinguish, irrespective of whether the combination was, for example, [action lexigram + demonstrative gesture] or a [goal lexigram + action gesture]; thus, his lexigram ordering superseded the ordering of argument functions (Greenfield and Savage-Rumbaugh 1990). While Kanzi thus showed an arrangement pattern that was not dependent on specific form–meaning pairings, it is characterized essentially only by one ‘‘rule’’ (Kako 1999: 8). Some patterns of linear order were also found in the gorilla Koko (Patterson 1978a: 91).
Regular patternings in the order of form–meaning pairings have been reported on a number of apes, and they tend to match word order in spoken English—the language used by the trainers. The chimpanzee Ally showed a preference for an order ‘‘demonstrative–noun’’ in 92 percent of his two-sign constructions and ‘‘subject–verb–object’’ order preference in 89 percent of his three-sign constructions (Miles 1978: 109–10), and in the two-sign utterances of the gorilla Koko, 75 percent of the 98 attributive phrases recorded had the order ‘‘adjective–noun’’ (Patterson 1978a: 91).
Sensitivity to linear arrangement was also found in the trained bottle nosed dolphins Phoenix and Ake, who could correctly enact relational sentences only by inferring the thematic roles of objects (transport vs. designation) from their syntactic position (Herman et al. 1984; Herman 1987). That they are sensitive to word order is suggested by the fact that they responded correctly above chance level to reversals of arguments when asked to distinguish between PIPE FETCH HOOP ‘‘Fetch the pipe to the hoop’’ and HOOP FETCH PIPE ‘‘Fetch the hoop to the pipe’’.
And there are also indications that the animal developed iconic ordering patterns. Savage-Rumbaugh and Lewin (1994: 161) observe that the action–action combinations that Kanzi developed were essentially iconic: The authors found that orders like tickle–bite and chase–hide were significantly more frequent than their inversions.
We mentioned earlier an interesting ordering distinction found in the orangutan Chantek: When the object referred to was present, then the form of an utterance would be object-GIVE, but when the object was not present, then the form was GIVE-object (Miles 1990: 519).
A number of the findings made in these studies suggest that linear arrangement is—at least to some extent—‘‘lexically’’ rather than ‘‘rule’’ based. Terrace (1983: 23) and associates found that their chimpanzee Nim used the sign more in the first position in 85 percent of the two-sign utterances in which more appeared (e.g. more banana, more tickle), 78 percent had give in the first position, and 83 percent of the instances in which a transitive verb (e.g. hug, tickle, give) was combined with me and Nim had the transitive verb in the first position. Note that Nim’s teachers had no reason to sign many of these combinations. That linear arrangement tends to be lexically based is an observation that has also been made of the language acquisition of children (e.g. Tomasello 2003b; Diessel 2005).
Analyzing the ninety-one sequences of signs that the chimpanzee Washoe, trained in ASL, was able to produce (Gardner and Gardner 1971), McNeill (1974) observes that this ape had nothing corresponding to syntactic patterns in human languages, but she displayed a constituent order where the addressee or party being addressed always preceded the addressor or speaker, and McNeill suggests that the animal showed a linguistic ability that contrasts with that of humans, namely one that emphasizes interpersonal or social interaction.
